ZiTokenizer 0.0.8

ZiTokenizer: tokenize world text as Zi

ZiTokenizer

Tokineze all languages text into Zi.

support 300+ languages from wikipedia, including global

use

• pip install ZiTokenizer

from ZiTokenizer.ZiTokenizer import ZiTokenizer

# use
tokenizer = ZiTokenizer()  
line = "'〇㎡[คุณจะจัดพิธีแต่งงานเมื่อไรคะัีิ์ื็ํึ]Ⅷpays-g[ran]d-blanc-élevé » (白高大夏國)😀熇'\x0000𧭏2022２０１９\U0010ffff"
indexs = tokenizer.encode(line)
tokens = tokenizer.decode(indexs)
line2=tokenizer.tokens2line(tokens)

# build
demo/unit.py

UnicodeTokenizer

basic tokeinzer

ZiCutter

汉字拆字

'瞼' -> ['⿰', '目', '僉']

ZiSegmenter

word => prefix + root + suffix

'modernbritishdo' -> ['mod--', 'er--', 'n--', 'british', '--do']

languages

default using "golabl" vocob, others from https://laohur.github.io/ZiTokenizer/index.html

tokenizer = ZiTokenizer(vocab_dir)